Balancing economic and ecological benefits for hydro-junction operation based on the ecological flow from the four major Chinese carps: a case study from Xinjiang River, China

The construction and utilization of the hydro-junctions would change the water flow and affect the hydrologic process required for the survival and reproduction of the aquatic organism in the river basin. To investigate the influence of the construction of the Jiepai (JP) and Bazizui hydro-junction (BZZ) on the hydrologic process in the Xinjiang River and coordinate the benefits between power generation and downstream ecological protection of the four major Chinese carps (FMCC), a one-dimensional hydrodynamic model is established to simulate the river hydrological regime of the spawning site before and after the construction of BZZ. Meanwhile, eleven ecological hydrologic parameters are used to evaluate the degree of hydrological changes for the spawning site caused by JP and BZZ, and four hydrological methods are applied to calculate the minimal and optimal ecological flows to satisfy the demands of FMCC. The multi-objective operation model combined with the non-dominated sorting genetic algorithm II(NSGA-II) is established to evaluate the relationship among the objective functions and design the ecological operation scheme of BZZ. The results indicate that the hydrologic changes in the spawning site has been moderately changed since the construction of the JP and would change more after the construction of BZZ. An obvious antagonistic relationship between electricity generating and ecological changes are presented in the simulation results, and the multiobjective operation model could clearly increase the comprehensive benefits under three typical years (dry, average, and abundant) by 20%, 19%, and 34% with the corresponding electricity generating lost by 4.9%, 3.5%, and 4.1%, respectively. This study is expected to provide scientific guidance to coordinate the restrictions and conflicts between economic and ecological benefits of hydro-junction operation in the lower reaches of the Xinjiang River.


Introduction
Hydro-junction is one of the most important tools for the utilization of water resources to meet flood control, power generation, and transportation needs (Chen et al 2016, Wu et al 2019. However, the impacts such as natural hydrologic rhythm change have attracted a growing awareness due to these impoundments on instream environment (Yang et al 2019). Especially, the variation range of water flow is limited to a small range and the number of high-flow pulses decrease, which would artificially lead the natural environment tend to be monotonous (Luo et al 2021a, Cao et al 2022. Based on the previous research, the ecological flow is fundamental to sustaining the aquatic environment and affect the availability and function of river habitat (Naiman et al 2008). Therefore, quantitatively evaluating the flow regime changes has been adopted as an effective measurement of river health (Gao et al 2018).

Study area and methods
2.1. Project overview 2.1.1. Study area The Xinjiang River is one of the five major rivers in Lake Poyang with a total length of 313 km and an area of 16890 km 2 from 28°59′ N to 28°44′ N and 118°05′E to 116°23′ E (figure 1) (Yuan and Forshay 2020). The average annual streamflow of this river is 182.28 × 10 8 m 3 , while the mean annual precipitation and temperature are 1586.4 mm and 17.8°C, respectively. Baita River, with a total of 161 km and a drainage area of 2839 km 2 , is the main downstream tributary of the Xinjiang River and originates from the Guangze city of Fujian Province. Based on the previous research, a total of 40 fish species from 10 families and 4 orders were found in the Yingtan section of the Xinjiang River, with 29 species in Cypryformes, which accounts for 72.5% of the total number of indigenous fishes (Chen 2010). The section of the spawning site is about 7.3 km away from the downstream BZZ and about 41.7 km away from the upstream Jiepai hydro-junction (JP), where the river way becomes more narrow and has a certain degree of slope and relief amplitude, which provides suitable conditions for the spawning of FMCC (Hu et al 2015).

Hydro-junctions in the Xinjiang River
Along the Xinjiang River, hydro-junctions including JP and BZZ projects have been built to meet the requirements of power generation and transportation. JP is the first hydro-junction in the lower reaches of the Xinjiang River, where impoundment was initiated in 2003 (Huang 2000). As presented in figure 1, the river is divided into east and west branches in Yugan County and empties into Lake Poyang. BZZ, located 49 km downstream from JP, began river closure in 2017 and is composed of Hushanju Hub (HSJ) in the east river and Mopilin hub (MPL) in the west river. The installed capacity and annual electricity generating of BZZ is 12.6 MW and 4312 × 10 4 kW·h, respectively, and the proposed scheduling scheme of which is shown in table 1.   Where t is the time coordinate (s); A is the cross-section water area (m 2 ); Q is the discharge (m 3 s −1 ); x is the range coordinate (m); q is the Lateral inflow single width flow rate (m 2 s −1 ); g is the acceleration of gravity (m 2 s −1 ); C is the Chezy coefficient; R is the Hydraulic radius of the cross-section. Four components called river network, cross-section, boundary, and hydrodynamic parameters are included in the framework of MIKE11 model. In the hydrodynamic model of the Xingjiang River, the river network is digitized by using the topographic map as the base map. The section data from different locations of the Xinjiang River are inserted in the cross-section file. Historical discharge data at JP and the water level data during the same periods of the east river and west river in the typical years are used as the upstream and downstream boundary conditions, respectively.

Model calibration and validation
Historical water level data at Meigang hydrological station in 1977 and 2005 are used to calibrate and validate the hydrodynamic model, respectively. The indicators including mean absolute error (MAE), average relative error (MRE), root-mean-square error (RMSE), and Nash-Sutcliffe efficiency coefficient (NSE) are selected to evaluate the validity of the model. The result show a satisfying agreement between simulated and observed water level values, which indicate that the model could accurately simulate the flow process of the Xinjiang River (figure 2).
2.3. Eco-hydrological rhythm and ecological flow 2.3.1. Indicators selecting The natural flow is the primary driving force of the composition, diversity, and distribution of lotic biota and especially play an irreplaceable role in stimulating the spawning of FMCC (Arthington et al 2006). In this study, indicators affecting fish spawning are selected to establish the ecohydrology indicator system to analyze the influence of the construction and operation of the BZZ and JP on the hydrologic regime of the spawning site downstream of the Xinjiang River. The indicator system is divided into three groups including (a) mean flow of

Calculation methods of ecological flow
The calculation methods of ecological flow are mainly divided into the hydrological method, hydraulics method, habitat method, and holistic method (Tharme 2003). Among these methods, the hydrologic methods have been widely used in areas lacking field habitat research. In this study, four common ecological flow calculation methods are selected to calculate the ecological flow of the lower reaches of the Xinjiang River. Then the calculation results are adjusted by adding flow pulse factors according to the spawning flow requirements of FMCC.
(1) Q90 method define the ecological flow by a discharge that corresponds to the 90th percentile in the flow frequency curve.
(2) Multi-year average flow method of the driest month define the ecological flow by selecting the average value of the measured flow data of the driest month from the long series of hydrological data.
(3) Texas method define the ecological flow by selecting the mean value of the measured discharge data in the driest month in the long series of hydrological data.
(4) NGPRP method define the ecological flow by selecting the flow of the normal flows year group with a 90% guarantee rate as the instream flow.

Operation model 2.4.1. Objective function
To maximize the comprehensive benefits of power generation and ecological conservation of BZZ, the objective functions are set to maximizing electricity generating, minimizing ecological changes, and maximizing the comprehensive benefits.
(1) Maximizing electricity generating Where F is the electricity generating during the total period, E i t , ( ) is the electricity generating of i-th hydro-junction during the period t, N i t , is generation output, K i is output efficiency, Q i t , is the outflow, H i t , is average water head, t D is the time step(s).
(2) Minimizing ecological changes The ecological benefits are evaluated by the lance distance: WhereD t is the lance distance between the outflow and ecological flow during the period t; Q 1 w is the lower limit of ecological flow, Q 2 w is the upper limit of ecological flow.
(3) Maximizing comprehensive benefits b and 2 b denote weights corresponding to objectives of electricity generating and ecological changes.

Constraints
(1) Water balance constraints: (2) Outflow constraints: ( ) are the minimum and maximum outflow of i-th hydro-junction during the period t, respectively.
(3) Water level constraints: (4) Output constraints: ( ) is the power output of i-th hydro-junction during the period t; N i t min , ( ) are the minimum and maximum power output, respectively.
(5) Minimum shipping flow constraints: ( ) is the minimum shipping flow of i-th hydro-junction during the period t.

Optimization algorithm
GA is a simulation method proposed by John Holland and characterized by a randomly generated initial population of a string of variables called chromosomes (Hu et al 2014). The Operation model is simulated by NSGA-II, which is improved by GA and has been broadly used to solve various optimization problems in many engineering and scientific contexts (Tsai et al 2015). Five operation steps that constitute the algorithm are setting parameters, generating the initial population, crossover and mutation, evaluation of the objective functions, and non-dominated sorting. The specific solution steps are as follows: (1) a random parent population is generated, then the offspring population is created based on crossover, mutation, and variation; (2) the objective functions are calculated for the parent and offspring populations and the combined population that includes parent and offspring is classified based on non-dominated sorting; (3) the crowding distance is computed for the combined population, and these population are sorted based on the crowding distance; (4) the optimized population is truncated to the same size as the parent population to generate an offspring population for the next iteration.  Table 3 summarize the ecohydrology indicators and the mean deviation of each indicator in the spawning site downstream of the Xinjiang River. Taking the data recorded from the period of post-BZZ as an example, the high-flow pulse duration have a high degree of change, while five indicators including April and May mean flow, low-flow pulse duration, high-flow pulse count, and rise rates have a moderate degree of change, and the rest of indicators have a low degree of changes. Taking the mean deviation as the degree of ecohydrology change, the average degree of ecohydrology change during the period of post-JP is 12.58%, while the degree during the period of post-BZZ is 36.14%, which indicate the overall ecohydrology changes in the spawning site are moderate during the period of post-JP and become severe during the period of post-BZZ. The first group of ecohydrology indicators reflecte the magnitude of the monthly mean discharge. Mean flow in the spawning periods decreasing during the period of post-BZZ compared with the natural flow values with obvious variations during April and May. The high-flow pulse is defined as the daily flow higher than the 75th percentile of daily flow from the natural condition, while the low-flow pulse is defined as the daily flow lower than the 25th percentile of daily flow from the same period (Gao et al 2018). Four indicators related to the flow pulses have decreased by an average of 47% with the greatest decrease in high flow pulse duration up to 75%. In addition, mean positive and negative differences between consecutive daily values have all decreased and get severe with the increase of hydro-junctions.

Ecological flow acquisition
Four hydrological methods are used to calculate the ecological flow based on the historical discharge data measured by the Meigang hydrological station for 65 years. Figure 3 show the ecological flow calculated by the NGPRP method is relatively high, while the values from the Texas method are generally low. In addition, the flood occur in April and June with the peak discharge are 527 m 3 s −1 and 544 m 3 s −1 , respectively. Compared to the mean flow of each month, the ecological flow process is basically consistent with the natural flow process. Furthermore, the ecological flow calculated by different hydrological methods is all less than the average annual discharge, which is in accordance with the characteristics of the ecological flow (Xing et al 2018).
FMCC are typical drifting egg-laying fish and the egg of them need a length of river characterized by a minimum turbulence or velocity to keep semi-buoyant (George et al 2015 ). Water temperature and flow rate are the major environmental factors that affect the reproduction and development of FMCC ( Table 4). The lowest water temperature for spawning is 18°C while the increasing flow is the signaling factor of spawning process. When water temperatures reach a suitable range, FMCC migrate to spawning grounds, in which they complete gonad development and begin spawning stimulated by high flow pulses and current events such as flooding . Therefore, the ecological flow including optimal, maximum, and minimum flow is obtained by referring to calculation results and considering factors such as the ecology and flow requirements of FMCC comprehensively, which is shown in table 5.

Simulation results of operation model 3.3.1. Relationship among three objectives
Based on the hydrological frequency analysis, the years 2011, 1987, and 1977 are selected as the dry, average, and abundant years, respectively. The inflow of the above years are used to set up simulated conditions, and the relationship between electricity generating and ecological changes based on different inflow of typical years is analyzed quantitatively. With the simulation results of the dry year as an example, feasible solutions from the per-iteration simulation results, which could satisfy all constraints with maximum comprehensive benefits are selected to analyze the relationships among the electricity generating, ecological changes, and comprehensive benefits. Figure 4(a) presents that the comprehensive benefits increase with increasing electricity generating and decreasing ecological changes. The electricity generating, ecological changes, and the comprehensive benefits range from 3850 × 10 4 kw·h to 4465 × 10 4 kw·h, 0.196 to 0.245, and −6.39 to 27.76, respectively. The gradient of ecological changes increases with electricity generating, whereas shows decrease trend with comprehensive benefits. Comprehensive benefits represent a trend of rising with the electricity generating and then falling when the electricity generating exceeds 4500 × 10 4 kw·h (figure 4(c)). Figure 4(d) represent that the comprehensive benefits show an increasing trend with the decrease in the degree of ecological changes. The relationship among the three objectives in average and abundant years are similar to that in the dry year. In the average year, the highest value of the electricity generating and comprehensive benefits are 4883 × 10 4 kw·h and 33.74, respectively. While the minimum value of ecological changes is 0.207. All simulation results in the average year statistics yield that the electricity generating, ecological changes, and the comprehensive benefits range from 3685 × 10 4 kw·h to 4883 × 10 4 kw·h, 0.206 to 0.267, and −36.37 to 33.74, respectively. In the abundant year, the highest value of the electricity generating and comprehensive benefits are 5324 × 10 4 kw·h and 37.06, respectively, and the minimum value of ecological changes is 0.235. All simulation results in the abundant year statistics yield that the electricity generating, ecological changes, and the comprehensive benefits range from 3879 × 10 4 kw·h to 5324 × 10 4 kw·h, 0.235 to 0.271, and −35.85 to 37.06, respectively. Figure 5 shows the results under both the original and ecological operation of the BZZ in three typical years. The inflow is obtained by observed values at the Meigang hydrological station in the dry, average, and abundant years, respectively. For the dry year, figure 5(a) shows that the ecological operation of HSJ and MPL moderately increases the flow magnitude in July and the dry season (from February to April) respectively to maintain the ecological flow requirements. For the average year, variations mainly occur in June from HSJ, May, September, and October from MPL. For the abundant years, the inflow is impounded in MPL around June and is increased in July to maintain the ecological flow.   Table 6 reveals that there has been a marked decrease in electricity generating and ecological changes from the abundant, average to the dry year. In contrast, this trend is reversed in comprehensive benefits. From table 6, the electricity generating under ecological operation is smaller than that of original operation, which could be attributed to the more ecological flow to mitigate the ecological changes. Compared with the original operation scheme, the electricity generating under dry, average, and abundant year with ecological operation decline by 4.9%, 3.5%, and 4.1%, respectively, while the ecological changes under three typical years with ecological operation decline by 7.9%, 6.3%, and 8.2%, respectively. With fewer economic loss, 25.2%, 19.2%, and 34.3% of more comprehensive benefits could be allocated to the Xingjiang River under the ecological operation strategy in dry, average, and abundant years, respectively.

Discussion
Indicators of hydrologic alteration could quantitatively assess the characteristics of hydrological changes, which has been widely applied for the research on hydrological regime change assessment, ecological environment impact assessment, and eco-environmental discharge estimation (Gao et al 2012, Souter 2017, Song et al 2018. In this study, the change of ecohydrology indicators reflecte that the construction of BZZ has further altered the  natural river rhythm of the spawning site in the lower reaches of Xinjiang. The cascade hydro-junctions play an important role in streamflow regulation and economic generation, while the operation scheme of which reduces the flood peak during the wet season and decreases the mean flow during the spawning periods. For the period of post-BZZ, the magnitude of monthly mean flow in the spawning periods is less than the pre-dam values, and the mean flow data of March and April are decreased by nearly 50%, which inevitably shorten the drift distance of fish eggs and affect the reproduction and growth of FMCC. Referring to the previous research, this decrease is mainly caused by the redistribution of the flow stored upstream of the reservoirs during the spawning period (Shao et al 2017). It has previously been observed that indicators including the frequency and duration of high and low flow pulses reflect no obvious changes in the period of post-JP compared with the values of pre-dam, similar to the results of this study (Liu et al 2016). However, no explanation for the effect caused by the construction of BZZ was provided. In this study, significant differences are observed in the indicators of the second group between the two periods of pre-dam and post-BZZ, and the post-BZZ values are all lower than the pre-dam values. Based on the previous research, flow pulses stimulated the ovulation and fertilization in FMCC, which accomplish gonadal development by stimulating high flow pulses or floods upstream (Guo et al 2011). With the high-flow pulse as an example, the mean values of frequency and duration of the flow event decreas by 43.35% and 74.6%, which reduces the stimulation signal for the spawning of FMCC.
Compared with the natural flow, the mean positive and negative differences between consecutive daily values decrease by 55.76% and 20.64% after the construction of BZZ, which limit the variation range of water flow to a small range and weaken the spawning signal of FMCC. No apparent change is observed in the number of hydrologic reversals, which is opposite to that of other research (Liu et al 2016). This is because the previous study has not considered the influence of BZZ, while it impound water with low frequency in the case of no  extreme floods happen. Based on the above analysis, the hydrological rhythm of the spawning site downstream of Xinjiang has been changed due to the construction of BZZ, which lay an impact on the suitability and availability of the spawning site for FMCC. However, besides hydrologic alterations, reservoirs construction could also cause effects such as blocking migration, release of hypolimnetic cold water, and habitat fragmentation, which have been shown to be the major drivers of fish species decline (Kitanishi et al 2012, Cheng et al 2015. To mitigate the antagonistic relationship between economic development and aquatic ecosystem protection, the hydro-junctions operation should consider the ecological demand. However, traditional hydrological methods are subjective and ignore the needs of the aquatic (Tan et al 2018). Therefore, the optimal range of ecological flow in this study is determined by considering the calculated results of ecological flow and spawning requirement on FMCC comprehensively. As shown in figure 3, the range of ecological flow is defined in the specific standard which is incorporated into the spawning demand of FMCC to reveal the degree of ecological changes. From the result of the above method, the ecological flow from March to May calculated by the NGPRP method acquire the highest values, so the ecological flow in the spawning periods is obtained using this method. To improve the spawn efficiency of FMCC, the optimal and minimum ecological flow from March to June is increased appropriately.
The calculated flow rhythm is consistent with the natural runoff, in agreement with the previous studies which have indicated that the ecological flow is calculated using the hydrological method from historical discharge data without consideration of the influence of human disturbance (Xing et al 2018). As shown in figure 3, flow from May to July accounts for the highest proportion of annual flow, and two flood processes appear in March to April, and May to June, respectively, which correspond to the aquatic organisms especially FMCC reproduction in the lower reaches of the Xingjiang River (Li et al 2019). For the calculate results of the ecological flow, the maximum flow could be utilized for drinking, irrigation, and electricity generating, while the minimum flow is designed to guarantee the basic need of the river ecosystem, and the optimal ecological flow values are between the maximum and minimum values.
The available study has demonstrated that there is an obvious antagonistic relationship between electricity generating and ecological changes, and the finding seems consistent with this study (Dai et al 2017). Previous research has been set maximizing power generation, flood control, or transportation optimized as objectives of reservoir operation with the basic ecological flow as the constraint, which could merely guarantee the minimum ecological requirement of the aquatic ecosystem (Jiang et al 2016). However, the comprehensive benefits present proposed in this study maintain the right balance between energy production and environmental protection. The simulation results in this study show that the smaller ecological changes with more electricity generating lead to higher comprehensive benefits. From the simulation results in different typical years, the distribution of ecological changes are more scattered in the abundant year, while they are more concentrated in the dry and average year, which indicate that reservoir inflow would make the scheduling of hydro-junctions more complicated and cause the simulation results to be more scattered.
From figure 5, in the dry year, the ecological operation scheme show a relative increase in flow during the spawning period while there is a relative decrease in the mean flow during the dry season, and the operation schemes are opposite in the average and abundant years. The reason for this phenomenon is that the inflow in the dry year during the spawning could not meet the ecological flow demand, so the inflow in January and February is stored to increase the outflow from March to June. The electricity generating is higher for the abundant year than those for the dry and average year in both the original and ecological operation scheme, while the situation is reversed in ecological changes, which means ecological operation could not balance the electricity generating, ecological changes. Although the antagonistic relationship between electricity generating and ecological changes are inevitable, this conflict could be mitigated as much as possible through the operation objective of comprehensive benefits. Multiple studies across the globe have demonstrated that the ecological operation could alleviate the negative effect of hydraulic projects such as hydropower stations, reservoirs, and hydro-junctions and this study corroborate the ideas of previous research (Zhou and Guo 2013. The hypothesis of this study is that the ecological changes under both the original and ecological operation follows the order of dry > average > abundant year. However, the result in table 6 suggests this hypothesis is incorrect, which is because the ecological changes mainly result from the fact the outflow of BZZ exceeds the maximum flow during the wet season. In addition, higher electricity generating would need more complex and unpredictable operation requirements, which lead to an increase in ecological changes. Since the construction of the cascade hydro-junctions in the lower reaches of the Xinjiang River, the sites where the flow regime is suitable for the spawning of FMCC have been damaged greatly. The ecological operation model of BZZ aims to quantitatively assess the response relationship between electricity generating and ecological changes, which is favorable for designing of the appropriate scheme to maintain the power generation while ensuring the ecological demand of the aquatic environment. In ecological operation, the ecological objectives have been integrated into the scheme with various functions coordinated, which could maintain the ecological environment while promoting benefits, eliminating harms, and realizing sustainable utilization of water resources (Dai et al 2017)

Conclusion
In this study, eleven indicators are selected to analyze the hydrological changes caused by BZZ in the spawning site. The multi-objective operation model based on the ecological flow is established to analyze the relationship among the operation objectives and design the ecological operation scheme of BZZ to balance economic and ecological benefits in the lower reaches of the Xinjiang River, enabling the following conclusions to be reached: (a) The hydrological regime changes in the spawning site downstream of the Xinjiang River are moderate during the period of post-JP and become severe after the construction of BZZ, which primarily manifeste in the decrease of mean flow and high-flow pulse. These changes inevitably reduce the stimulation signal for the spawning of FMCC.
(b) The rhythm of ecological flow calculated by the hydrological methods is consistent with the natural flow, and the flood occur in April and June, respectively. Based on the spawning demand of FMCC, the ecological flow from March to June is increased appropriately to improve the spawning efficiency. (d) Compared with the original operation, the comprehensive benefits under three typical years with ecological operation improve by 25.2%, 19.2%, and 34.3% at the expense of electricity generating lost by 4.9%, 3.5%, and 4.1%, respectively, indicating that the ecological operation is beneficial to the ecosystem.
(e) The ecological operation model proposed in this study has focused on the balance of economic and ecological benefits based on the ecological flow by the FMCC. In the next step, further studies are needed to comprehensively consider the affect such as blocking migration, release of hypolimnetic cold water, and habitat fragmentation caused by the construction of reservoirs, and carry out more systematic research on the ecological operation of the cascade reservoirs.
would be published elsewhere. All funding sources supporting the work and the institutional or corporate affiliations of the authors have been acknowledged in the title paper.